The Future Of Autonomous Mobile Robots

Mateusz Sadowski, Robotics and Drone Consultant at M.Sadowski Consulting, and Mike Oitzman, Founder of Mobile Robot Guide, join Jeff Dance to discuss the future of autonomous mobile robots (AMRs) to replace dangerous and monotonous jobs that don’t fit with humans. They also dive into the exponential growth of AMRs in the market due to cost reduction and robotics as a service model, fleets of robots being the next step on the evolutionary curve for mobile robot companies, and their vision of the future of autonomous mobile robots ten to twenty years from now.

Mike – 00:00:01: Companies automating the movement of grapes from those field row workers to the packout station. And the workers love it because they can just sit there and pack as fast as they can throw it on, and they’re getting credit for much higher volume and the rows get picked faster. So it’s a win win for everybody.

Jeff – 00:00:19: Welcome to The Future Of, a podcast by Fresh Consulting, where we discuss and learn about the future of different industries, markets, and technology verticals. Together, we’ll chat with leaders and experts in the field and discuss how we can shape the future, human, experience. I’m your host. Jeff Dance. In this episode on The Future Of, Autonomous Mobile Robots, we’re joined by Mike Oitzman, founder of the Mobile Robot Guide and a Robot Report contributor, and Mateusz Sadowski, an Engineer and Robot Consultant passionate about Autonomous Mobile Robotics, also with deep experience in Avionic Robotics. Welcome. It’s a pleasure to have you both here on the show.

Mike – 00:01:05: Great, thanks for having us.

Mateusz – 00:01:07: Yeah, thank you.

Jeff – 00:01:08: Awesome. Mike, if we can start with you. Can you tell the listeners a little bit more about your experience in the space and why you’re passionate about AMRs?

Mike – 00:01:17: Well, thanks, Jeff. Yeah, so I came to this industry three plus decades ago, starting as an applications engineer for Adept Technology and then ultimately found my way into product management, spent some time in several of the leading enterprise software companies in Silicon Valley, and then ultimately got recruited to come back to Adept in 2015 to run product management for their mobile robot product line. And I was super excited about that, primarily because I had been become a coach for a FIRST Robotics team when my son was in high school. And so I got passionate about mobile robots again and the opportunity to get deeper into the industry. I jumped on that. I was at Adept for multiple years until the organization was sold to Omron and at that point became a consultant. Tried to figure out what I wanted to do, but I had all this information in my head about the market for mobile robots. So I started the Mobile Robot Guide as a way for the industry to benefit from that knowledge, and then two years ago sold the publication and joined the WTWH Media family as a part of the editorial team for The Robot Report and now The Mobile Robot Guide as well.

Jeff – 00:02:30: Awesome. Thanks for that background, Matt. Tell us more about yourself.

Mateusz – 00:02:34: Thanks, Jeff. So I’m a Robotics Consultant primarily. I’ve been consulting for five years now, and I started working in robotics around eight years ago. I’m mostly focused on mobile robotics. However, the last year or two or two, I’ve been working with a company that’s in heavy metal industry. So the company is called DTE and they are making robots that shoot lasers into metal alloys to analyze the metal composition. So it’s really interesting. Quite far away from mobile robotics, but every now and then, I still work with companies that look into implementing some software for robotics platforms. As a consultant, I’ve been doing quite a lot of review of what’s going on in robotics, and this also sparked my side project Weekly Robotics, a newsletter where every week I deliver news and information about interesting projects in the industry.

Jeff – 00:03:37: Awesome. Thanks for your background. Before we nerd out on autonomous mobile robots and I love the topic, we’ve also been working with AMRs at Fresh for over five years. What do you guys do for fun?

Mike – 00:03:51: Well, I’m a photographer, so I love taking pictures of the night sky, the Milky Way. I love being out. It’s so peaceful to be in the evenings looking at the night sky. And so that’s what I love doing when I’m not sitting at the computer writing about robots.

Jeff – 00:04:09: Nice. Matt, what about you?

Mateusz – 00:04:11: It’s kind of similar, Mike. Although I’m not a photographer, but I love being in nature, and I really love hiking. So when I can, I try to be out in the mountains and just walking and enjoying. Other than that, I have to say my biggest problem is that my work is my hobby, so I need to work on that a bit more to diversify, I think.

Jeff – 00:04:36: Yeah, I think as engineers, as people in the robot space, it’s good to have a balance. And I think we’re so connected to technology. How can we disconnect? So getting out in nature and taking photos sounds like a great way to remind ourselves that we’re still human. Good. Well, let’s just start with some basic one-on-one, and then in the present day, and then let’s evolve into the future of autonomous mobile robots. There’s a lot happening in the space, and that’s one of the reasons we wanted to do this episode, because there’s so much progression that’s happened in the last few years, and things have been moving really fast. Let’s start with what is an AMR and kind of what is the difference between an AMR and an AGV? Mike, can you start with kind of helping the audience understand that?

Mike – 00:05:21: Sure, Jeff, I’d love to, and I explain this all the time to folks who are considering what it means to be an autonomous mobile robot. And really, industrial robots have been around for decades, and really, one of the big differences with an autonomous mobile robot is that it runs around on its own power and has to do all of its decision-making on board and perceiving the environment and running its tasks. And there is a technology that’s been around for, I would say, probably three or four decades now, which is called Autonomous Guided Vehicles, and they’ve been working primarily in the early days of automotive. And an AGV is the forerunner of an autonomous mobile robot, or AMR. And really, one way to think about the difference is that AGVs have historically been deployed for tasks that move material around a factory, I equate it to sort of like running a bus route throughout the factory. Whether they’re moving material, like auto bodies as the car is being built, or plane wings, those type some AGVs have been huge mechanisms, but they run a predefined path, much like a bus route would run around the factory, maybe even dropping off material or picking up finished goods, that type of thing. So they stay in their lane. They do avoid obstacles. So if there’s a box or a cart or a vehicle or a human, they won’t run into it because they know it’s there, but they don’t go around it. They stop and wait until it either moves or they send an alarm and somebody has to come recover the vehicle, get it back on its path, what have you. And in the last, probably since around the early 2000s, were the first implementations of autonomous mobile robots. The idea that a vehicle that could run around without being constrained to a specific path, find their own way, run their own sort of missions or tasks from point to point, and I would equate that to a taxi or an Uber. Right. The idea that these vehicles could be called to go to someplace in the factory or in the warehouse or on the street, pick up something, and then find their own way to an endpoint where they do something deliberate or what have you. And if something gets on their way, they can either stop and wait for that to move and then be on their way or plan an alternate path to find their way to their destination. And in a warehouse, that might mean backing up and going down a different aisle, one that’s not occupied by a Fork Truck, for example. And that’s one of, I think, the significant differences between an AGV and an AMR technology today is that the AMRs can make their own decisions. They can repass, plan where they’re going, find a different way without having to call an operator to come recover them from some situation. That’s not to say that AGVs aren’t inexpensive because they still may be expensive devices because of the types of things that they’re doing, like carrying an airplane wing. That’s an expensive vehicle. But in terms of the intelligence that’s on board, the AMRs that we see today being developed today are much more intelligent, doing much more of the compute on board than an AGV has ever done in the past.

Jeff – 00:08:27: Thank you. To be autonomous, you need all that sophistication versus just to be guided. Is it fair to say that the AMRs tend to be smaller, smaller vehicles doing lighter tasks, and AGVs tend to be larger in general. 

Mike – 00:08:39: I think, but primary prescription of being able to maneuver, that they tend to be deployed as smaller vehicles. And one of the things you said the guided vehicles were guided by things like magnetic tape or painted lines on the floor. That was the primary guidance, was following that sort of path. And so AMRs today are being deployed in very small packages.

Jeff – 00:09:06: In addition to the conversation we had with our guests on today’s episode, we asked another expert to provide their insights on the future.

Harshil – 00:09:16: My name is Harshil Shah and I am a Robotics Engineer at Fresh Consulting. There are several reasons why AMRs have seen an exponential growth in recent years and why this trend is likely to continue in the future. Technological advancements in available computing resources, sensors, and algorithms have enabled AMRs to navigate in dynamic environments and perform tasks autonomously and efficiently. The cost of AMRs is still slightly on the higher side, but it has reduced significantly compared to a few years ago, enabling a quicker ROI and providing a stronger business case. The boom in ecommerce has led the supply chain industry to look into more efficient fulfillment solutions, therefore increasing the demand for AMRs and warehouses, fulfillment centers, and other logistics settings. Labor shortages, paired with increased demand has accelerated the deployment of robots that perform those tasks. The benefits from improved safety and increased productivity have been the major drivers towards rapid adoption of robotic solutions.

Jeff – 00:10:24: The market for AMRs is, I think, around a few billion right now, but it’s supposed to grow to 10 billion in the next five years. Matt, why do you think there is a lot of exponential growth right now with AMRs?

Mateusz – 00:10:38: I think pandemic could be one of the reasons. We had many people not in the workforce and companies were investing heavily into automating some of the jobs, and I think this could be one of the reasons. And the other, if I had to guess, that could be connected with the workforce availability is the price. I believe the price of these robots is going down. And also with some models like robotics as a service, they can be attractive for even smaller company that before would not consider getting a full robot for themselves.

Jeff – 00:11:22: Yeah, just to give one example of that, I think the LiDAR components that are often used with autonomous mobile robots, I think we were paying, like, 30,000 dollars for some of the Velodyne pucks, like, five years ago. And now I think we can pick up some stuff from lots of companies for maybe $300 that are much smaller, but there’s lots of options. That’d be one example of a key technology that is probably driving that price to be lower, and I think that matters. Mike, do you have any other thoughts on some of the exponential growth and we anticipate in the next five years?

Mike – 00:11:52: I want to follow on the thread that you just described, which is the availability of sensors. And we’ve seen that with the anticipated growth of autonomous vehicles, the sensors that are going on autonomous vehicles, that’s one of the markets that’s helped bring down sensor costs, the advent of cell phones because they have helped to miniaturize the compute technology, made it way more powerful. Again, it continues to get more powerful with every new generation of chip companies like Nvidia and their platforms, Qualcomm and their platforms again finding other applications. So we saw Nvidia come over to robotics from their experience in gaming, right? So gaming, which is a hugely popular market, a huge market, took the reapplication of GPUs from gaming applications to now processing AI and machine vision models. And so that’s all helped improve what robotic engineers like Matt can do to do more powerful things on board the actual device. And so that just opens up more applications now, I think, for the robots. And that’s really been the explosion, literally in the last five years. The types of vehicles, the number of vehicles, the number of configurations that are in the marketplace today.

Jeff – 00:13:11: Great. As we talk about that marketplace and that explosion, what are some of the most common applications for AMRs?

Mateusz – 00:13:17: I would say Pick and Place comes to mind. Locus Robotics, who is one of the unicorns, I think the only AMR unicorn. They specifically do that. And I think recently they announced that they did 1 billion picks with their robots. Is that right? Do you guys know?

Mike – 00:13:36: Yes.

Mateusz – 00:13:37: That’s quite an achievement. I think.

Mike – 00:13:39: So, Jeff I’m tracking over 1100 different vehicle configurations in my database today, and this is individual products from different companies. So in that, I try to categorize each vehicle by the type of application that it’s designed for. And in that database, north of 20% of those vehicles are for logistics. So, as Matt said, moving something from one place to another, that could be fork trucks, that could be picking up boxes, pallets, that type of stuff. And then on top of that, another, I think about 8% of the market is warehouse specific applications. This would be specific warehouse workflows, goods to person type applications, person to good type applications, a variety of different workflows. So that by far is the biggest market segment that I see, that I’m tracking. And then right after that would be vehicles that I just described as platforms. Bare Metal Mobile Robots that don’t have a specific application payload on top of them or solution they’re designed for integrators or for companies to take and then put their own type of payload on top of that robot to do something like whether that’s shelving whether that’s a lift mechanism, whether it’s a specific flow cart, cart moving all those types of applications in a variety of different applications, where you want to start with the technology and you want to do something on top of it. So that’s sort of the second biggest segment. And then the third biggest segment that I’m super excited about, because I just got back from the World Agriculture Expo last week in Central California, is agriculture. Agriculture is just blooming with, pardon the pun. With agricultural robotics, there’s so many ways that they’re being deployed today. So many new companies attempting to build platforms for a variety of different applications, from cultivation to harvesting. And that’s after the big two tractor companies, in case New Holland and John Deere, which have already automated a variety of their platforms, you got a bunch of other smaller players that are also attempting to bring applications. So that’s another market segment that I’m excited about.

Jeff – 00:15:45: Just as a micro economy, given our business, we’ve actually worked in those three industries. And so to hear those, it resonates from a little micro perspective because we actually do custom robotics and custom workflows for companies. And so to kind of hear that from you and kind of all that research, that definitely resonates. As we think about you started to mention some companies. You mentioned Locus. You mentioned platforms like Clearpath and MIRROR kind of came to mind. But what are some of the big companies you think that are in the AMR industry now? You mentioned Locus Robotics as sort of one of the unicorns. Who are some of the other big players?

Mike – 00:16:27: Well, I think you have to put Amazon in this. Amazon started with the acquisition of Kiva Robotics more than a decade ago, at that point for the unicorn. But since that time, Amazon has built a huge internal robotics division. So they don’t sell their products to other companies, but they are designing and deploying those solutions now inside of all of their warehouses. So the technology that they’re advancing, I think, is noteworthy, even if we don’t see it on the general market. And I think a couple of other names that come to mind. There are some competitors now to companies like Fetch, including companies like GreyOrange, and Geek+ who are deploying large solutions. So with the advent of Amazon taking Kiva off the market, that opened up the opportunity for companies like GreyOrange, Geek, Berkshire, Grey, and a variety of other similar companies to build similar solutions for warehouses that are not Amazon based. And I think that’s where that market is growing.

Jeff – 00:17:32: As we think about, Matt, you mentioned the Pandemic and sort of we’ve seen a lot of supply chain issues through that pandemic. Things got shut down, our economies got disrupted, and people weren’t in their normal routines. We also saw that the advent of so much more ecommerce. So how has that sort of affected AMRs? Mike, can you give us more your perspective on that?

Mike – 00:18:00: I sort of equate it to the growth in this warehousing solution comes at a variety of application needs. And I think when we saw during the Pandemic, I saw this personally, the fact that more folks were shopping online, getting deliveries to their home. And of course, that helped companies like Amazon and other ecommerce companies to bring those types of products, expanded their portfolio of home products. People couldn’t shop out in the public. But the other market that has grown in this time frame has been grocery. So before the Pandemic, the grocery market was very small. Only a few select groceries were attempting to do home delivery of products. Maybe big retailers like Walmart had begun to do some of this, ASDA in Europe and a few others. But I think what the Pandemic did was it took this concept of the urban, maybe millennial population who were ordering food at home and ordering groceries at home and expanded that to other generations, older generations who were not used to doing that and now saw a real opportunity to get stuff delivered. And so that meant that, for example, grocery had to figure out how to do curbside delivery or to do home delivery and then to also automate the actual picking of food. And I think there we’ve seen growth in another side, not just the mobile robots, but in something called Automated Storage and Retrieval Systems, or ASRS systems, these huge highly dense storage arrays. That product is stored in bins and then you have these tunneling robots that go pull the bins out to find the specific bin with chips or potatoes or whatever the case may be and to fulfill orders. And so you’ve got a lot of large companies in that market have also grown with robotic automation as a part of their solution as well. And there’s a lot of interesting things that are being done in that market segment as well that we don’t see because it’s all happening inside the walls of a large warehouse.

Jeff – 00:20:11: It’s good to hear that from your perspective, given your expertise and focus here, Matt, from a technology perspective, my understanding is that you work heavily in ROS 1, ROS 2 and also the NAV stack. How have those advancements kind of affected AMRs?

Mateusz – 00:20:28: Yes, I think ROS 2 is a major shift when it comes to software on robotics. Before we had this ROS 1 that also, I think, dramatically changed the industry because suddenly we had people developing code and they were willing to share this code. So as a company looking to develop software for robotics, you suddenly didn’t have to write this driver for the sense for each of your sensors, but you could just pull it from the Internet and it’s almost like plug and play. And because the standards are so defined, it just worked. Now ROS 2 introduces some different concepts, especially when it comes to introducing some different concepts, especially on the messaging side, for example. So you don’t have this central entity called ROS master that tells each of your software nodes where it needs to send the messages and who needs to receive them. Instead, its communication is process-to-process. So it changes quite a bit because your systems can be more distributed. The learning curve for software developers is a bit steeper, but at the same time you can create more robust systems. And one thing that came with ROS 2 is the Navigation 2 stack, which is some of the concepts are based on what we had in ROS 1. But it’s I would say on another level because suddenly we have these Behavior Trees where we can define very flexible system and you can decide what needs to happen in instances where your robot cannot get somewhere or they have another issue. And I think this flexibility will be huge for many people that look into just the navigation side of things.


Mike – 00:22:29: There’s so many parallels, Matt, to the laptop development, the computer development over the last couple of decades as well, right? The fact that you could throw in any component, hard drive or what have you in the laptop as you configured it. I think that ROS makes that a bigger opportunity for new developers who have an idea they can get quickly started and get further down the road with much less investment to prototype and then ultimately productize their solution. So what’s your experience that side of productizing a solution?

Mateusz – 00:23:00: I agree it feels like magic sometimes because I’m doing some side projects, for example, and then I know I have this sensor and this sensor and I just need to make them work. And sometimes I’m surprised that it’s just, I just pull this software in and then it just works with the tools that I have. As long as I stick to some defined practices, I can combine the RTK GPS and a lighter sensor and suddenly I have a 3D scanner. It won’t be perfect because it’s software development. There needs to be some certain steps, but it’s good enough for me to capture the map of some buildings or at least the facade of some buildings and build a point up from that. And if I can do myself something like this over the weekend, I think it’s huge because there are companies that’s been specializing in that and maybe my solution will work 80% of the time, but I build it over the weekend. So I think it’s hugely powerful. And this in turn allows you to experiment quite a bit. So your RND can be quite fast, you can evaluate if some sensors work for you. And yeah, I can talk and talk about ROS, but I think the thing that I love the most is that I can record the data and then I can replay it. And this is huge for any I call it passive part of the system. So like you have SLAM, for example, you have data that’s coming in and the system is not reactive to SLAM because things are moving. But your SLAM is not navigation, so it doesn’t expect the robot to turn or do any action. And because of that I can just record the data and then I can tune and tune-in very fast iterations so I don’t need to pull out the robots anymore, go to the field and waste time. I can just record one good set of data and just work with that and this also enables me to be a remote consultant for companies. So I’m very glad for these kind of tools.

Jeff – 00:25:22: Just to elaborate on that a little bit, you mentioned RTK GPS, which kind of helps you know where you are, right? You mentioned SLAM from a mapping perspective to kind of help map your local environment. We talked about LiDAR to kind of detect things that are close to you. You mentioned Behavioral Tree in ROS2 and how that’s advancing. As you think about we talked about some of these core components that are advancing and making it easier to plug and play together so that we can accelerate all sorts of different solutions. Mike, you mentioned hey, you’re tracking thousands of these different vehicles that have a lot of these core components in them. And just for the audience, I think that now that we’ve introduced some terms here, I think as you think about use-cases, we say, hey, I need to go pick something up. I need to go pick something, move something, pick something, hand something over. The Behavioral Tree allows you to see, like, hey, is the elevator open or not? Is the door open or not? Is this thing in my hand or not? Has it been given to me or not? And you start combining these things, and then the use-cases, when these things are done, well, explode right in the office. There’s a thousand use-cases in the warehouse, there’s hundreds of use-cases. But you need all these kind of core components to come together. And it seems like the infrastructure is sort of here and it’s still developing. Some of the problems I see in this space are related to integration. And Mike, I’m curious, from your perspective, you said you’re tracking like thousands of different vehicles. Do you see anyone leading in that space as far as integrating one robot with another robot or being aware, like, oh, this is how I communicate with this other robot?

Mike – 00:27:00: Yes. In fact, this concept of bringing together fleets of robot is one of the next steps on the evolutionary curve for any mobile robot company. You start with a single robot, can it do the single task, pick up something, move it here, what happened, what have you, but as soon as you want to replicate that deployed in a real factory, a real warehouse, a real restaurant, or on the street, or you have to deal now with multiple robots doing things. For example, in retail, we’ve seen robots that do inventory management now that do floor scrubbing, that do security. Right? And so we’ve had some very high visibility stories. I’m not going to replay those for us here. But the fact is this last year that trying to deploy robots from a variety of different vendors into a retail store means that you’ve got a different task mission interface for each of those robots. Each of those robots have different chargers. The whole charging problem is being resolved in a variety of ways with wireless charging. Now universal charging stations. But the fact is, of those thousand vehicles that are in my database, there’s 1000 different charging solutions, each of which have a custom connector to charge that specific robot. So there has been no standardization in terms of physical charging. There’s becoming some standardization, for example, in wireless charging. Now in several companies that have come to market, Wiferion and Robotic, the other two big leaders are a couple of other companies are also doing this. Now, CAPAL is another Israeli company we just talked about, reported on getting some money this week, also coming into this market. But the idea there is to create a universal charging interface. And now you need to work with those robots to use that device, because again, you may not have enough chargers for every robot if you’re going to keep them working all the time. So they have to figure out how and when to charge. So those are just some of the tasks that are important as we up level that to how do you communicate with all those robots? One of the interesting things that’s happened in the last 18 months is that MassRobotics out of Boston chartered an interoperability committee. A number of the leading robotic suppliers have participated in that. And the whole idea was to just say, hey, can we start with just reporting our positions? Right here’s where I am in the world. They’re not even sharing maps yet. It’s like, but here’s where I am in the world. Here’s a reference point for that. So they all have a reference point. Let’s report that so that if I meet somebody in the Nile, I know it’s another robot and we can make a decision to stop, wait or whatever is going to happen there. But the fact is, that was the first thing. The other thing that they did was to put chargers onto the map so that if a charging station was a universal charging station, then one robot could say, I’m now occupying this charger, so it’s not available. So other robots could make their own decisions about whether and when to go charge. And I think the next level of progress will be to create some universal operating system. There are some companies out there that are doing software for this space, and this includes companies like Formant, Melly, Move.ai, to name a few. So they are trying to create some universal software for this. And so first today they have to create adapters or connectors to all of these various robots. But at some point there will be some universal language in which I think things communicate and that’ll simplify, as Matt said, the integration of all of these things.

Harshil – 00:30:44: Several principles come to mind when thinking about designing robots for the future. Safety should be the top priority to prevent accidents, for protecting people and infrastructure assets that surround them. AMR should be adaptable to be able to navigate new environments and changing conditions. Efficiency in terms of energy consumption, time to complete a mission and task dispatching across the fleet to improve resource allocation and utilization should be kept in mind. AMR should be designed to be scalable. It should be able to operate within a larger network of heterogeneous robot fleets, machinery and IoT devices. AMR should be designed to be accessible by anyone, regardless of their technical and physical abilities.

Jeff – 00:31:34: Let’s shift and talk a little bit more about the future. We think about the future of autonomous mobile robots. What do you guys envision 10, even 20 years from now? Matt, start with you.

Mateusz – 00:31:45: Can actually a problem. When I think about these robots in warehouses, I often imagine a problem. So you have someone driving a forklift and there’s something that’s not right as a person. If there is a driver, it’s quite easy to stop them. You just wave your hands, you show, and we have the universal language. With the robots, it’s a bit more tricky. And I think even recently there were cases where firefighters were breaking autonomous vehicle glass to stop them because otherwise they wouldn’t stop and they would drive over the fire hoses. And this shows that we don’t have a way really to communicate with these robots as humans. You can probably hope that it detects you as an obstacle, but it could be a safety risk, especially if we try to move faster and so on. So I would be really keen to see robots interpreting us as humans better. I think this is a core problem to solve in the near future. And recently also you hear about ChatGPT quite a bit, and I don’t know if these kind of models could be useful, but maybe it will be the first step to communicate a bit better.


Jeff – 00:33:10: So communication with humans is a key kind of problem to keep working on as we envision the future. Using different AI models that can try to interpret our behavior and make sure things are safe is an important aspect of integrating robots in with us as humans as we think about the future of work, maybe the future of the home. Mike, what are your thoughts as we envision the future 10 to 20 years from now? What do you see?

Mike – 00:33:38: Well, I think robots are going to continue to find their way into other places that they’re not today. That’s an easy answer. The difficulty is going to be in making them safe. One of the areas that I think has seen quite a few starts and stops and nobody’s quite got it right yet, is in this concept of social robots putting robots into the home. The long promised Rosie, the robot who can pick up your house. We’ve got robot vacuum cleaners today, and I’ve got one that’s very intelligent, does a great job of cleaning the floor, but it doesn’t pick up socks. It doesn’t pick up dog toys, right? It doesn’t move things around, it doesn’t vacuum the couch and pick up all the pet hair, that type of stuff, as well as it doesn’t do the dishes. And that’s just me as an able-bodied person, let alone helping. We’ve got aging population around the world in the coming two decades, and I think that, again, we talk about the limited folks to actually do jobs. Support for that elderly community is going to be a gap between caregivers and the needs of people. So I think that’s one of the areas we’ve got to solve and will be solved with some innovation. I don’t know what that innovation is yet, but I can tell you today that the systems that are out there are not quite crossing that chasm yet. There have been some interesting solutions by, for example, a company called Labrador Solutions that has a mobile robot that can move trays. It can pick up a tray, deliver a tray to somebody. There’s another company called Hello Robot that’s created a prototype. A great laboratory solution still has some way to go before it becomes feasible for deployment, but those are the early solutions. I also had the chance, Jeff, to visit Carnegie Mellon last year, visit several labs at Carnegie Mellon, where there’s some research being done in these areas of helping in the kitchen, helping around the house, opening doors, opening drawers, teaching robots to do those simple tasks. As Matt said, being able to look at something and understand what its form and function is just by identifying what the thing is that the robot is looking at. Is it the scissors? Is it a knife? Is it a telephone? What is it? There’s so much opportunity for research and advancement in this next decade that it’s hard to say what the solutions are going to look like. But I know where the needs are, and I think that’s one of the areas where there’s a tremendous need to support folks, elderly folks, to support folks who might be handicapped or have a disability, being more able. So there’s lots of opportunities there for robots, and operating in the house is a tremendously difficult environment. You’ve got stairs, you’ve got furniture, you got all kinds of things way more complex and complicated than working in a factory or a warehouse where things are defined and the workflows are a lot more obvious and easy to define. That’s why that’s been the low hanging fruit for the last couple of decades for the deployment of these solutions.

Jeff – 00:36:50: I would agree with you that the home is not yet tapped, and yet there’s so much promise there. And we know iRobot the biggest sort of robot company, if we call them a robot, and Amazon just bought them. So we would anticipate some major investment sort of advancement there, given the platform that they have, a lot of the components that they have. I also noticed that Dyson has been investing, I think, more than a billion dollars into kind of home robotics and trying to build on top of their product line. So it’d be interesting to kind of watch them. But for the most part, I don’t think social robots. I mean, I was at the World Expo in Dubai and saw some interesting things with some social robots and their intent to I don’t think most of them have failed. I would say so far the consumer robotics hasn’t been successful, but we know there’s still serious problems. You brought up the aging population and there’s lots of innovation, I think, happening in Japan related to robotics and then AMRs being like the robots that are moving around, but then also have the ability to kind of help. And I think one of the things I’ve been really interested in is sort of the notion of the mobile robots with an arm attached. As you think about those two components, when we attach an arm to a robot or something else that can manipulate or connect or attach, then all these sort of use-cases sort of come up versus just material handling, like moving things around, which is still a key asset. We’re constantly moving things around in warehouses, in the construction sites, and maybe not so much in the home. Maybe we want some things brought to us. But I think when you attach the arm or you have the AMR go to an arm, then all these really interesting use cases kind of come up. As we think about the future, what sort of is the next big disruption that helps it be even more common in businesses or in the home?

Mateusz – 00:38:48: I think when it comes to disruption, these robots need to earn money ultimately, so they need to be priced lower than the human equivalent. Let’s be honest, that’s how market works. And I think if the price of compute and the sensors goes down, it hopefully will enable this. But general intelligence is very hard and I think this is something that will be quite tricky to solve and I don’t know how we get to that point. And building these universal robots is also quite a challenge, I think. I wish I had the answer, but I think the first company that solves that so makes this kind of robot will be quite successful. But yeah, I don’t know what I think we just need to keep going and see where we get to.

Mike – 00:39:45: I agree with Matt in terms of there’s a lot of technology still to evolve going forward. And again, mobile robots are self-contained solutions. So part of the problem on the technology side that I hear from a lot of companies still remains power, battery power, because they’re battery powered. So as batteries improve, we’re getting so many more electrical things, as more electrical vehicles come into market, that the demand for battery technology evolution is going to support this market as well. Giving us longer running solutions and more power to do things like compute. So edge computing is the other area that needs to evolve and we’re seeing advancements there as well because perception is hard. It takes a lot of electrons to run those engines that do the perception. Matt talked about all the sensors that might be on a given platform. They all take energy as well. And again, a trade-off that any robotics engineer has to make when you have limited power is how many sensors can I use and what’s the least power hungry device that I can put on this? So I get the longest runtime to fit the application. Either that or I have to again figure out how and when to recharge or how I get energy back into the system. So I think those are some of the things that we’re going to continue to see benefit from adjacent markets. The biggest, I think, disappointment that everybody has seen is we expected we’d have autonomous vehicles running on the road by this time. That promise has been there for ten years. I think we’re still ten years away from a level five capable vehicle in your driveway. I think we’re going to see robotaxis be the application because again, those fleets are owned by a company, not by people, and they’re going to continue to be expensive. But over time, that autonomy is going to find its way into the vehicles that we drive every day in the form of safety systems, ADAS systems. And we can benefit from that in the mobile robots market because all of those things are going to cut the cost of those sensors and improve the reliability and what have you. And Jeff, I just want to come back to and say, looking forward again, I’m very bullish on this agriculture market. I see so much innovation coming from other areas, both from industrial robots as well as mobile robots, and again, the AG market. We’ve got to feed a growing world population and do it effectively and efficiently and repeatable. And not only does that mean that the labor involved in picking, harvesting, cultivating and what have you, but you see systems now that are optimizing the use of chemicals and what have you, or even reducing the need for chemicals or eliminating the need for chemicals which provides a much safer food stream for humans and pollinators and all the rest. So there’s an amazing amount of innovation that’s happening there and I’m really looking forward to following that market growth over the next decade. And again, it becomes one more huge application segment for all of these sensors, all of the compute that’s going to benefit everywhere else that these things are being deployed.

Jeff – 00:42:55: These all make sense to me. As far as disruptions that need to keep evolving such that this space can keep advancing, I would add the compatibility just like as we think about doing more complex tasks, how do you harmonize one robot to another robot and make that communication really seamless. Obviously lots happening there, but it seems like we’re early stages. It’s an area that we’re pretty passionate about at Fresh, as we think about robots, I think humans often fear robots, but the reality is we feared the computer when it came out and we thought it was going to completely change and work and actually did. It did change work completely but didn’t replace us. It augmented our capabilities and helped us automate a lot of things. Often as we think about robots, the dirty, dull, dangerous, monotonous sort of tasks have been targeted because they don’t fit as well with human beings. Would you say AMRs are similar? Do you see things that are kind of more in the dirty, dole, dangerous, monotonous sort of space being the areas where things will really grow first?

Mateusz – 00:44:09: I would agree that AMRs will replace some of these jobs that are monotonous, at least. Again, picking tasks come to mind and no person enjoys, at least I expect not many people enjoy walking 20km a day putting things in the box and just sending it away. So I think it’s a huge chance for this.

Jeff – 00:44:35: Those jobs exist. But I think in the pandemic when people had an opportunity to reflect, these aren’t jobs that people came back to. The industries where the greatest demand is for labor are the jobs that not as many people came back to. And I think people had more opportunity to reflect. As a human being, like, what do I care about most, what’s meaningful to me? And it wasn’t the monotonous jobs where automation could play a role. So I think that was one of the takeaways that the entire global industry took. And I think that’s why we’re seeing a lot more investment in the space.

Mateusz – 00:45:11: I agree. And one company I’m following in particular called Greenzie, they make Autonomous Mowers. And I know that in the US there’s huge issues with labor shortage for people that can mow loans. And they automate this sector and it seems like they choose quite a nice area to build expertise in.

Jeff – 00:45:35: I think that’s a good example. Mike, I bet you’re tracking a lot of companies in the Autonomous Mowing space.

Mike – 00:45:41: We are, and I think that’s another market segment that I’m excited about Greenzie. SKY is another good company here in the US that’s making trends. But just to come back to your point, Jeff about dolan dangerous. I think we’ve proven this. Or it’s been proven that. Applications for mobile robots, we come back to Locus, for example, they are automating that long haul between the pick aisle and the packout station in the warehouse, so that the pickers do what humans do best, which is selecting an individual item out of a bin and putting it into a box. But no one wants to do the long walk with the box to the packout station right in the warehouse. And so the humans are doing less walking. They’re being held interactive all the time because they just sit there back and forth pulling stuff. It’s maybe a little bit more interesting than the long walk part of that. And they’re not as tired at the end of the day, so there’s not the burnout. So there’s all of that. Another example, again, that I learned about last week at the World Ag Expo, a company that’s automating for table grape harvesting. So, again, they can’t figure out how to harvest table grapes because they’re so fragile, and they have to snip off bad grapes right away when they’re picked so that they don’t ruin the bunch. So the field workers snip the grapes off, they put them in boxes, and they get paid by the number of pounds that they pull in a given day. So those workers are inspired to pick as many grapes as fast as they can. And they hated the walk of two, three, four, five hundred yards to take those full boxes to the infield packing station, where they’re then sorting everything, weighing it, getting it ready to go to the store. That’s how that works. And so this company is automating the movement of grapes from those field row workers to the packout station. That’s all those little mobile robots do. And the workers love it because they’re making more money, because they can just sit there and pack as fast as they can throw it on, and they’re getting credit for much higher volume, and the rows get picked faster. So it’s a win win for everybody in that sense.

Harshil – 00:47:54: Bill Gates famous quote that we always overestimate the change that lucker in the next two years and underestimate the change that will occur in the next ten resonates with me. Predicting the future is challenging, but an educated guess can be made. Based on the current trends and advancements in the pipeline, AMRs are likely to have increased intelligence and autonomy that will enable them to perform more complex tasks. There will be continued enhancement in safety. One example could be better predictive collision avoidance algorithms, which will enable robots to operate at higher speeds, leading to increased productivity, enhanced collaboration with humans to perform tasks and assist with day to day operations. There is a need for interoperability standards and its adoption to enable management of heterogeneous fleets of AMRs and integrating with other technologies such as IoT devices. With the reduction of cost over time, robots will become more accessible to small businesses and individuals.

Jeff – 00:49:00: As we close out, where can we learn more? Mike, you have some serious involvement in kind of looking at the industry. Tell us more about we can learn more about, you know, AMRs and sort of the world of robotics.

Mike – 00:49:11: Well, thanks, Jeff. Of course. I’m a journalist, editor. Write every day about what’s happening in the world of robotics on the Robot Report and on the Mobile Robot Guide, and also a podcast host for the Mobile Robot Report podcast. But the other thing that I would love to invite folks to is we’ve got two big events coming up this year. We’re trying to create some premiere events for robotic developers. The Robot Summit in Boston in May 10th and 11th is a great opportunity for anybody that’s developing robots AMRs or arms or any of the technology stack around that. We’ll have a lot of great conference discussions. There are some great keynotes already on board for that. So I invite you to come join us there. And then again, we’ll do it again in October on the West Coast of the US in Santa Clara with the RoboBusiness event, one of the premier robotics industry networking events. And if you’re a startup, it’s also a great place to learn and meet and network with folks who are investing in robotics. We bring a lot of investors there, a lot of high level executives there from a variety of companies just exchanging ideas. And also we are running the second of our new event called the Field Robotics Engineering Forum. So I’m sharing that event, and again, why I’m so passionate about what’s happening in the field, in AG and mining and construction, just to name a few of the things that are happening outside of the buildings that we’re going to be focusing on at that event. And that happens in mid-October in Santa Clara RoboBusiness.

Jeff – 00:50:50: Thanks, Mike. Matt, for you, what are some of the go-to places for learning more or going deeper? And that could even be technical resources you’re a fan of?

Mateusz – 00:50:59: Yeah. When it comes to learning about new things, I really like the Robot Report, so thanks, Mike, for being part of this. I also enjoy articles in IEEE Spectrum. They also do very high-quality articles and then I think prime example of technical writing about robotics perception is Tangram Vision. I’ve never seen a company so devoted to their blog. So if anyone is looking to create high-quality resources, please use them as an example and then let me know so that I can include you in the newsletter, because I think this really helps our industry, like having all this information that’s high quality and that we can use, especially as engineers.

Jeff – 00:51:49: Thank you. Last question. What excites you most about the future of autonomous mobile robots?

Mike – 00:51:55: I love learning about all of the different implementations of mobile robots and discovering that That’s why I love my job today. Having been in high tech, been on the sell side the majority of my career, now I get to sit in the middle. Everybody gets to tell me their exciting stories about or their crazy ideas for the next market changing robot. So that’s what I love. I love to hear those stories. I love to retell those stories. I love to amplify that knowledge and information. So I’m excited about that. I’m also have to say I’m excited as a first mentor for robotic competition, working with the next generation of engineers. I love inspiring those kids to a career in engineering, different careers in engineering beyond what they might have thought they wanted to do in learning and playing with robots. It’s so much fun to watch the excitement and watch the lights come on and to follow their careers as they grow and join the ranks of working to build the next set of technology.

Mateusz – 00:52:56: I’m really excited about the growth of robotics, and I think Boston Dynamics is a great example here. If you watch the videos from five years ago and from today, it’s a huge change. And every time they release a new video where they take the robots to robotics, to the next level, I feel like a kid during Christmas. It’s a good present, and I’m also similarly excited to how the software develops. And with ROS and ROS 2 being implemented by many players, I hope the ecosystem grows and we see more and more advanced applications. And as a final thing, I would really love grocery delivery robot to deliver me some package at some point. I think this will be quite exciting when it happens. We don’t have so many of these in Europe, unfortunately.

Jeff – 00:53:54: I don’t know if you saw that Amazon shut down their Scout division, but we’ve been working in that space for a couple of years. But it seems like that last mile edge, there’s a lot of companies still working in that space, some that we frequently communicate with, and I think that’s a huge space still. Nonetheless, thank you for your insights and wisdom. As engineers, as leaders in the autonomous mobile robot space, as consultants, it’s been great to get your perspective. Really excited about where this space is going and appreciate your perspective on the future.

Mike – 00:54:29: Thanks, Jeff.

Mateusz – 00:54:31: Thank you, guys. It was nice having a chat with you.

OUTRO – 00:54:35: The Future Of Podcast is brought to you by Fresh Consulting. To find out more about how we pair design and technology together to shape the future, visit us at freshconsulting.com. Make sure to search for The Future Of, on Apple podcast, Spotify, Google podcast, or anywhere else podcasts are found. Make sure to click subscribe so you don’t miss any of our future episodes. And on behalf of our team here at Fresh, thank you for listening.